VENDORiQ: Google’s Botched Launch of Bard Reveals a Bigger Problem

The Latest

14 February 2023: Google fumbled its AI search service’s launch (really, pre-launch). The Bard tool is powered by the company’s conversation technology LaMDA (Language Model for Dialogue Applications) and is touted to offer high-quality responses. However, during a presentation last week, Bard made some mistakes, which caused investors to sell off the search company, cutting the value of Google’s parent company, Alphabet, by more than USD$100 billion.

Why It’s Important

The frenzy over generative AI – ChatGPT, Dall-e, Midjourney Bing Search, and Google Bard, was predictable and lamentable. Generative products have been available for some years now, though in specialised areas such as high-volume social copywriting, search engine optimisation (SEO) copywriting and graphical design for advertising.  

IBRS has long predicted that AI will continue its rapid growth in many industries, given the various use cases that it is relevant. The current ‘cultural shock’ of AI is partly a result of a poor understanding by managers of how technology evolves. 

As IBRS outlined in a 2017 presentation on the future of work, technology advances at a predictable rate that, at specific intervals (plus or minus a few years), results in tipping points in terms of performance-to-cost ratios for specific applications. Generative AI is very computationally intensive, but with vast Cloud services and the world’s digital knowledge and conversations freely available now on the internet, making AI cost-effective for general use (as opposed to costly niche applications) is such a tipping point.

The irony here is that even major technology companies such as Google and IBM (remember Watson?) can be swept up in the hype – all because they failed to recognise (largely predictable) patterns of technology evolution.

For Google, the launch of Bard is a humiliating and costly example of falling into the hype.  Google has a range of AI (as does Microsoft) that will be just as transformative as the current frenzy over ChatGPT. 

One example of Google’s power in AI is Google BERT (Bidirectional Encoder Representations from Transformers). BERT is a pre-trained language model developed for natural language processing tasks such as question answering and sentiment analysis. It uses an attention mechanism to learn contextual relations between words (or sub-words) in a sentence. The model is pre-trained on a large corpus of text, allowing it to be fine-tuned for specific NLP tasks with smaller amounts of labelled data. Sound familiar? OpenAI’s GPT (Generative Pretrained Transformer), and Facebook’s RoBERTa (Robustly Optimized BERT Pretraining Approach) are similar approaches to generative text. All have also achieved state-of-the-art results on various NLP tasks and are commonly used in industry and academia.

However, Google chose to present a ‘me-too’ product in what looks like a panic over losing ground in search. What Google has fallen into is the myth that ‘first movers have the advantage’, whereas history tells us, ‘second and third movers get the money in the long run’.

Who’s Impacted

• The entire C-Suite: 
• CIO / CTO
• Business strategists
• Data teams

What’s Next?

• Start briefing your entire C-Suite. First, ensure they understand this AI is a predictable evolution, and it’s only going to get bigger now the cost-performance tipping point has been reached.  
• Help senior executives understand the depth (and speed) at which AI-powered products will take over the manual and ‘creative’ (really, semi-creative) tasks.
• Explore how to leverage AI tools built into existing customer and help desk service solutions. 
• Before adopting any AI-powered solution, consider methods to measure user satisfaction and segment the results to find specific areas to improve in the future.
• Pay particular attention to privacy and the potential impact of algorithmic bias. 
• Ensure that your organisation has identified its key SaaS platforms or vendors and has understood its specific roadmap for the delivery of AI functionality.
• Don’t be like Google. Don’t rush. Don’t get scared by the hype. Take a steady, tested and well-evaluated approach to leveraging AI.

Related IBRS Advisory

1. DIY or ready-made? Choose your AI adoption path carefully

2. Artificial Intelligence Digital Assistants are rescuing Customer Service